Background: Starting from 1 January 2007, electronic locking devices based on proof-of-age (via electronic cash cards or a European driving licence) were installed in approximately 500 000 vending machines across Germany to restrict the purchase of cigarettes to those over the age of 16.
Objective: To examine changes in the number of tobacco vending machines before and after the introduction of these new measures.
Design: The total number of commercial tobacco sources in 2 selected districts (70 000 inhabitants) in Cologne were recorded and mapped. This major German city was the ideal setting for this study as investigators were able to use existing sociogeographical data from the area. A complete inventory was compiled in autumn 2005 and 2007. A total of 780 students aged 12 to 15 were also interviewed in the study areas. The main outcome measures were quantities and locations of commercial tobacco sources.
Results: Between 2005 and 2007 the total number of tobacco sources decreased from 315 to 277 within the study area. Although the most obvious reduction was detected in the number of outdoor vending machines (−48%), the number of indoor vending machines also decreased by 8%. Adolescents changed from vending machines to other sources for cigarettes, particularly kiosks or friends (+31% points usage rate, p<0.001; +35% points usage rate, p<0.001, respectively).
Conclusions: Although the number of tobacco vending machines decreased, this has not had a significant impact on cigarette acquisition by underage smokers as they were able to circumvent this new security measure in several different ways.
Statistics from Altmetric.com
Until 2007 Germany was considered “paradise” for adolescent smokers. More than 700 000 vending machines, or 1 machine per 16 adolescents, facilitated access to cigarettes virtually at any time.1 2 In particular, those mounted outdoors in public spaces (eg, near schools, playgrounds, sports facilities) provided unhindered and anonymous access to cigarettes. It was not uncommon for children as young as 11 or 12 years old to purchase cigarettes from vending machines, despite the restriction on smoking for those under the age of 16 in place since decades.3 Apart from its significance as a source of tobacco, the presence of vending machines within public spaces also normalised the perception of cigarettes and facilitated their acceptance. Additionally, tobacco vending machines in Germany are frequently found mounted next to gumball or candy vending machines, further increasing their appeal (fig 1).
Recently, an attempt to regulate tobacco sales in Germany to underage children has been undertaken in line with Article 16 of the World Health Organization Framework Convention on Tobacco Control,4 signed by the Federal Republic of Germany in 2003 and ratified in 2004. Article 16 advocates for two basic options concerning tobacco vending machines: limit access for adolescents or enact a total ban on tobacco vending machines. Germany opted for the first alternative, thus on January 1, 2007, a change in the Youth Protection Law (Jugendschutzgesetz (Article 1055 JuSchG)) was introduced.5 This law stipulates the mandatory installation of electronic security devices on all vending machines in Germany. With the electronic security device, consumers are required to insert some form of electronic identification (eg, electronic cash card or a European driving licence) to purchase cigarettes from vending machines. The security device only permits the purchase of cigarettes by vending machine users above 16 years of age. Furthermore, on 1 January 2009 the age limit was raised to 18 years and the machines reprogrammed accordingly.
The introduction of the electronic security devices has met with strong opposition from the tobacco industry, which had been very successful in resisting the complete ban of tobacco vending machines for several decades. As early as 1997, the German Association of Tobacco Wholesalers and Vending Machine Distributors agreed to a “voluntary self-restriction” proposed by the German Federal Ministry of Health. This restriction prevented the installation of vending machines and tobacco advertisements within 50 and 100 m, respectively, of schools and youth centres.6 Unfortunately, this has done little to affect access to cigarettes by those under 16 years.7 Similarly, the introduction of the electronic devices is largely considered by national smoking prevention experts to be another concession by the tobacco industry to avoid or delay a complete ban on tobacco vending machines.8 Germany is a country which continues to permit the use of tobacco vending machines in public areas; 22 other European countries have all banned these machines in outdoor spaces.9
To date, the effects of this new regulation on the supply of and demand for tobacco are not known, in part due to the lack of official data as well as the highly restrictive information policies of the tobacco industry and vending machine suppliers. Thus, it is not known what effects this law has had either on the number of vending machines available or the behaviour of adolescents who smoke. Nevertheless, this information is essential for the evaluation of such a measure, particularly since it is an implementation just completed or currently being considered in many other nations (such as Japan, Australia and Scotland).10
Thus, the aim of this study was to compare the number of tobacco vending machines and other commercial sources before and after the enactment of the new law in a geographically defined area. This study also assessed changes in acquisition habits within the relevant age group. Specifically, the research questions were: “Has the total number of tobacco vending machines and other commercial sources been changed after the implementation of the new law?” (primary research question), and “What changes could be observed in the behaviour of consumers under age 16 after the implementation of the new law?” (secondary research question).
Geographic Information System (GIS) study
Data for this longitudinal observational study were collected in autumn 2005 and 2007. The total number of the commercial tobacco sources (vending machines and other sources) within the study areas and period was recorded and plotted using a GIS.11 12
We carried out the Sources of Tobacco for Pupils (SToP) study in 2008 in selected districts of Cologne. Located in the western part of Germany, Cologne is one of Germany’s four “megacities”, with a population of 1 024 000 inhabitants. Cologne is generally considered to be a “typical German city” with a historical city centre surrounded by suburbs and districts. In Cologne the proportion of persons aged 0 to 25 years was 24.3% and the unemployment rate was 13.1% (compared with 25.8% and 11.7%, respectively for Germany). For this study, the two districts of Cologne–Muelheim and Cologne–Lindenthal were chosen.
The boundaries of districts and social areas have been delineated by the statistical department of Cologne13 “Social areas” are defined as conjoined clusters of homogenous streets, houses and quarters within a city, which share certain characteristics and attributes. Cologne was chosen for this study because it is one of the few German cities where these types of “social areas” have already been assessed with relevant status indicators. Indicators include, for example, income, percentage of immigrants, percentage of people receiving social welfare and the number of adults and children per housing unit. Heymann13 has described this procedure in detail. The city of Cologne contains a total of 269 of these “social areas”. The seven social areas of Cologne–Muelheim exhibit above-average unemployment rates (6% to 13%). Additionally, a greater proportion of the population in these areas receives social assistance (5% to 19%). By contrast, in the six social areas of Cologne–Lindenthal, the unemployment rates (1% to 3%) and proportion of the population receiving social assistance (0% to 1%) lie below the average figures for the city of Cologne. The same also holds true for additional indicators of status such as population density and motor vehicles concentration. In short, the district of Cologne–Muelheim is ranked as below average whereas Cologne–Lindenthal is classified as above average according to selected status indicators. We also considered the spatial dimensions of the two selected districts as well as their comparable distance from the city centre. Cologne–Muelheim covers an area of 7.7 km2, contains 3265 residential buildings, 21 301 households and 40 805 inhabitants. Cologne–Lindenthal covers 7.1 km2 and has 3497 residential buildings and 30 031 inhabitants.14 Both districts are located approximately the same distance from the city centre (4.3 km and 6.1 km for Cologne–Muelheim and Cologne–Lindenthal, respectively).
Outcome and data collection
The locations of tobacco vending machines and other commercial sources are not officially registered in Germany. Therefore, all commercial sources were mapped and geocoded by two geographers that inspected the study areas twice (2005 by CM and 2007 by DS). The first inspection was carried out in October 2005, followed by the second in October 2007. During the data collection process, all streets and squares within the defined study area were covered on foot or bicycle. Information pertaining to the sale of cigarettes in kiosks was obtained verbally. Additionally, other locations (eg, supermarkets, drugstores and gas stations) were checked for tobacco products. Inquires were made if these not on display in the immediate area of the checkout counters. All exterior (eg, wall mounted) cigarette vending machines, as well as those located inside service establishments (eg, fast food restaurants) were recorded. In those cases where either new points-of-sale had been installed or former ones removed between 2005 and 2007, the staff of the establishments were interviewed to confirm these developments. This was done to check for any missed point of sale in 2005 and thus exclude any potential mismatches in the recording process.
Method of geocoding
Mapping and geocoding was accomplished using the GIS software ESRI ArcMap (ESRI, Redlands, California, USA: ArcGIS V. 9.2). All recorded individual points-of-sale were first plotted then transferred to the GIS. To transfer the mapped results, we used digital details obtained from the Deutsche Grundkarte DGK5, scale 1:5000 (German National Maps, Berlin, Germany) as raster data.
We also conducted a student survey in both districts. The aim was to obtain information about any changes in demand that may have occurred as a result of the implementation of the new law. The survey was carried out anonymously using standardised questionnaires and was conducted between December 2007 and January 2008.
The questionnaire was developed in collaboration with three experienced schoolteachers. It was subsequently pretested with 68 students in the city of Karlsruhe, Germany. The Department of Youth, Children and Family Welfare (“Amt für Kinder, Jugendliche und Familie”) of the City of Cologne provided a list of all 11 schools within the study area. Information about the study was provided to these schools and their participation sought. A total of four schools participated: one out of the three secondary modern schools (Hauptschulen), one out of the three secondary high schools (Realschulen) and two out of the five academic high schools (Gymnasien) in these districts.
Inclusion criteria for the survey were age between 12 and 15 years and informed consent of the school director, the teachers and the participating student. As per the school law of Northrine–Westphalia (Article 120(4)), neither consent of the students’ parents nor ethical approval was required. A total of 824 students from 32 classes from grades 7 to 9 participated. Approximately 95% of the students (780) were below the age of 16. The overall response rate was 82.8%.
Outcome and data collection
All of the students were asked to provide their current smoking status. In order to validate this self-assessment, we asked all participants, including non-smokers, to estimate the proportion of smokers within their class (ie, “How many smokers are in your class?”). The validity of the self-assessment could thus be controlled by comparisons with the mean values obtained from the proxy assessments. The concordance of the two values was then calculated and estimated as the interclass coefficient (ICC).15
The smokers were also asked if the introduction of the electronic locking devices requiring age verification restricted their ability to obtain tobacco, their reaction to this development and the sources they were currently using to acquire cigarettes. All students who reported being smokers in 2005 were also asked to retrospectively recall their sources for tobacco during that year. The sources included in the survey included commercial and social sources (eg, cigarette vending machines, kiosks, supermarkets, petrol stations, friends, siblings and parents). The statistics software program SPSS V. 16.0 (SPSS, Chicago Illinois, USA) was used for all analyses.
In 2005, we counted 315 commercial tobacco supply sources within the 2 study districts, which is approximately 4.4 commercial sources per 1000 inhabitants. In total, 58 were outdoor and 126 indoor cigarette vending machines. By 2007, the number of commercial sources had declined to 277. Of these, 30 were outdoor and 116 indoor cigarette vending machines.
Additional tobacco sources included kiosks (n2005 = 78; n2007 = 72), supermarkets (n2005 = 23; n2007 = 24) and others (n2005 = 30; n2007 = 35). The overall reduction in the number commercial sources (−12.1%) resulted mainly from the removal of nearly half of the outdoor cigarette vending machines since implementation of the age control law (−48.3%). The total number of indoor machines also decreased but to a less significant extent (−7.9%).
In 2005 as well as in 2007, we found significantly fewer commercial tobacco sources in the socioeconomically privileged district of Lindenthal than the socioeconomically disadvantaged district of Muelheim. This can be seen in the number of sources per 1000 inhabitants. The number of commercial tobacco sources decresed from 5.3 in 2005 to 4.9 in 2007 in Muelheim. In Lindenthal, the number of commercial sources was 3.2% (2005) and 2.5 (2007) per 1000 inhabitants, respectively. This difference is particularly appreciable when the number of cigarette vending machines is compared (fig 2). Another difference between the two districts was that the reduction in commercial tobacco sources in Muelheim was primarily from decreases in the number of outdoor vending machines, whereas the reductions in Lindenthal were observed across all types of sources.
The spatial distribution of the commercial tobacco sources is shown in the GIS-based maps (figs 3A and 4A). They illustrate the almost ubiquitous availability of cigarettes in all areas. Any arbitrarily chosen point inside the investigation area was situated within a few hundred metres of a commercial tobacco source. Furthermore, it is clear from the maps that the outdoor cigarette vending machines are easily accessible from schools in both districts. No student surveyed during 2005 and 2007 had to walk more than 400 m from their school to reach an outdoor cigarette vending machine (figs 3B and 4B).
Finally, an almost uninterrupted chain of tobacco supply sources can be observed along main roads. In particular, indoor cigarette vending machines in both districts are concentrated along these routes. Indoor cigarette vending machines are often mounted in pubs and fast food restaurants, which are often located along major roads (figs 3C and 4C). Further analyses yielded no evidence of increased cigarette vending machines in the proximity of schools. Longitudinal analyses also did not suggest that more machines were either removed or mounted in areas near schools compared to other areas within Muelheim or Lindenthal.
Along with these changes in supply, there were also noticeable changes in the reported behaviour of adolescent smokers. In 2007, 9% of those aged 11 to 15 years old reported smoking. Another 29% reported having occasionally smoked in the past but had quit by the time the survey was conducted. Students who reported that they were regular smokers by 2005 were asked for their reactions to the new electronic locking devices. Three quarters of the current smokers claimed to have found a way of bypassing the electronic locking devices (fig 5).
For example, some solicited older schoolmates or friends to buy cigarettes from cigarette vending machines for them or borrowed the cash cards or driver’s licences of others. A total of 28% of these smokers reported using several strategies simultaneously (fig 6).
Several conclusions were drawn from the students’ statements concerning tobacco supply sources: Firstly, although commercial sources have lost importance for adolescent consumers, approximately 25% of underage smokers were able to bypass the electronic locking devices. Secondly, adolescents have increasingly resorted to kiosks and supermarkets to purchase their tobacco products since the implementation of the electronic locking devices. Thirdly, underage smokers are able to obtain cigarettes through friends, siblings and parents. Cigarettes are often taken unknowingly from packages at home purchased by parents.
Statement of principal findings
After the implementation of an electronic, card-based age control system in January 2007, nearly half of the existing outdoor vending machines were removed in the study areas (change in supply). The manner by which underage smokers obtain cigarettes has also shifted (change in demand). As the number of other commercial tobacco sources such as kiosks, drug stores, petrol stations and tobacco stores has largely remained almost constant since 2005, underage smokers have increasingly turned to these as well as social sources (mostly older schoolmates and friends) to obtain cigarettes.
Limitations and strengths
The methodical limitations of this study concern mainly the selection of the study area, the validity of the student survey and the establishment of a causal relation. Strengths of the study include the use of historically unique data, the collection of the data within the setting of a “natural experiment”, the longitudinal design, the inclusion of innovative GIS methods to visualise the number, distribution and density of tobacco supply sources as well as the simultaneous examination of changes in supply and demand.
Selectivity of the area of investigation
Although we have termed Cologne as a “typical German megacity”, it cannot be considered as representative of all German cities. The two selected districts share several common features including population, size of the geographical area and distance to the city centre, yet there are differences. We specifically incorporated the different social structures of the two districts into the study, which provided a key basis for comparison. Since there are no similar existing studies (neither for Germany nor any other country, as far as the authors are aware), this study was created to provide a detailed overview of the intended and unintended changes after the introduction of measures of this type.
Validity of the student survey
The selection of the schools for the student survey may also have affected the validity of the results. Schools that participated did so voluntarily, which may have introduced an element of self-selection bias. Nevertheless, the composition of the students in our sample matches the student profile in schools in Muelheim and Lindenthal (30% non-German students, 20% enrolled in secondary modern school, 30% in secondary high school and 51% in academic high school, compared with 29%, 10%, 28% and 62%, respectively, in our sample).14 Additionally, our figures of never (62%), former (29%) and current smokers (9%) correspond well with federal data for 12–15-year-olds for the year 2006 (62% never smokers, 27% former smokers and 11% current smokers).16 The proportion of smokers in Germany is slightly higher than that of Cologne, which is located in the western part of Germany (formerly the Federal Republic of Germany). This is the result of the difference in the smoking rate between the western and the eastern (formerly the German Democratic Republic) regions of Germany.16 Finally, the response rate of the students in these schools was more than 80%. This combined with a high ICC of 0.852 (p<0.001) strongly indicates agreement between the self-assessment and the proxy assessment.15
In addition to that, further calculations show that the composition of the sample matches the officially available structural data on pupils in this area:14 30% of the pupils within the entire investigation area are of foreign origin, in our sample this group comes to 29%. The total of the students within the investigation distributes to the different school forms as follows: 20% attend “Hauptschulen” (secondary modern schools), 30% attend “Realschulen” (secondary high schools) and 51% attend “Gymnasien“ (academic high schools).14 In our sample coincides roughly with these values: 10% attend “Hauptschulen”, 28% attend “Realschulen” and 62% attend “Gymnasien”.
It must be noted, however, that retrospectively asking the students about their cigarette-purchasing behaviour 2 years prior might have introduced recall bias. Additionally, a reduction in the sample size occurred as the pre/post analysis showed in fig 5. This is due to the exclusion of older students who were 14 and 15 in 2005 that were not included in the follow-up as well as those that were below the age of 12 at the time the first survey was conducted.
Establishment of a causal relation
The aim of this article was not to establish a causal link between the introduction of electronic locking devices and changes in adolescent smoking behaviour. For this, a pre/post survey of students of the same age would have been necessary in order to evaluate the effects of this intervention by age (ie, intraindividual purchase patterns between 2005 and 2007). Nonetheless, Robinson et al has recently demonstrated surprisingly time-stable source and access patterns among adolescents.17
Data show a continuous decrease in the smoking prevalence among German adolescents between the years 2001 and 2007.18 During the period of the study, several measures including tobacco tax increases, media campaigns and school intervention programs were implemented simultaneously. Thus, this prevents the direct linkage of the introduction of the electronic locking device to changes in patterns of adolescent tobacco procurement.
The German Association of Tobacco Wholesalers and Vending Machine Distributors (“Bundesverband Deutscher Tabakwaren-Groβhändler und Automatenaufsteller e.V.” (BDTA)) has stated that the high costs associated with the installation of the electronic locking devices have been the major factor in the reduction of the number of vending machines. According to the BDTA, the continual rise in tobacco taxes as well as increases in cigarette smuggling over the last few decades in Germany have also contributed to the reduction.19 Between 2001 and 2006 the total number of vending machines was between 700 000 to 800 000. This number fell to 430 000 units shortly after the conversion in 2007 and again to 420 000 in 2008.19 20 (Bundesverband Deutscher Tabakwaren-Großhändler und Automatenaufsteller, personal communication).
This study is the first to map commercial tobacco sources using a GIS, which provided information not only about the number but also the spatial distribution and density of sources. We were also able to combine this geographical data with information about underage smoking behaviour. The collection of data in 2005 presented us with the unique opportunity to longitudinally investigate changes before and after the introduction of the Youth Protection Law. The comparatively extensive area of investigation and the thoroughness of the data collection are further strengths of this study. In Germany (as well as in many other countries), neither the number nor the location of commercial tobacco sources are systematically recorded thus there is a need for this type of information. This study provides important evidence concerning changes in supply and demand after the introduction of these measures.
Positioning in the current state of research
As previously mentioned, the use of GIS mapping to plot commercial tobacco sources is unique to this study. Previous socioenvironmental studies have used GIS to study the locations of tobacco advertisements.21 22 For example, Hackbarth et al show that tobacco advertisements are more strongly promulgated in districts that are of lower socioeconomic status than in ones that are of higher status.21 Another finding from Luke et al was that tobacco advertisements are concentrated along major roads in St Louis, USA.22 The results of our study also agree with these findings. We also observed a significant, positive association between the unemployment rate within specific “social areas” and the number of cigarette vending machines (r = 0.545; p<0.05).23 Similarly, we also found that the concentration of indoor cigarette vending machines mounted in pubs and restaurants was higher along main roads. An advantage of this study is that we were able to map and analyse trends for all commercial tobacco sources for two districts and combine this with sociodemographic data and reported adolescent smoking behaviour.
According to studies that have examined the tobacco purchasing behaviour of adolescents, there are three typical reaction phases that occur after the introduction of cigarette vending machine restrictions.24 25 The first phase involves a significant decrease in machine-based purchases.25–27 This is usually followed by the discovery of ways around such blockades.24 28 29 Our survey data also show that underage smokers find ways around these restrictions by asking older schoolmates to purchase cigarettes for them or to lend them their electronic identification cards. Thus, 25% of all smokers in this study under the age of 16 were able to continue to buy cigarettes from vending machines.
In the second phase, adolescent smokers gradually resort to other commercial sources.24 The reason for this development is that age control measures are often easier to circumvent. As an example, test buys showed that in the 1990s the purchase of tobacco products was possible for underage consumers because of insufficient controls in stores in the US.25 28 30–33 After the implementation of stricter controls, the trend for underage smokers in the US was to target small stores and only very sporadically purchase from vending machines or large chain stores to bypass age control restrictions.17 25–27 The switch to smaller shops, particularly kiosks, was also observed in Germany after the implementation of the Youth Protection Law. Unlike major store chains, the proportion of tobacco sold is often more significant in smaller stores, which makes them much more sensitive to restrictions on tobacco sales. This may indirectly encourage proprietors to avoid questioning underage customers.
During the third phase, as the restrictions and controls facing adolescent smokers increases, a switch to social sources17 24 like older friends, siblings and parents occurs. Klonoff et al and Shive et al report that many adults do not regard this behaviour as problematic and willingly provide cigarettes to underage smokers.34 35 This type of behaviour was also observed in our study.
Even though the number of vending machines has been significantly reduced, many adolescents continue to evade age controls by switching to other commercial as well as social sources. This supports the conclusion that electronic locking devices on cigarette vending machines alone are not effective. Thus, we support the calls of international experts in favour of a complete prohibition of cigarette vending machines.1 8
What this paper adds
To date, it has been unclear how the total number of tobacco vending machines changes after the installation of electronic age control devices. This is the first study to have recorded and compared the number and pattern of vending machine locations before and after the implementation of a regulation of this kind.
After the implementation of an electronic, card-based age control system in January 2007, nearly half of the existing outdoor vending machines were removed in the German study areas. Nonetheless, many adolescents evade age controls by switching to other commercial as well as social sources.
Without the wealth of statistical information provided by the personnel of the “Office for Youth Support” in the “Department of Youth, Children and Family Welfare”, namely H-K Heymann and M Langenbach, this study would not have been possible. The authors would also like to thank Christina Huy and Silke Roehrig for their helpful comments and support in the preparation of this manuscript.
Competing interests: None.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.